In situ SERS reveals the route regulation mechanism mediated by bimetallic alloy nanocatalysts for the catalytic hydrogenation reaction

Synthesizing arylamines with high selectivity via hydrogenation of nitroaromatics is a long-standing challenge because of the complex reaction pathways. Revealing the route regulation mechanism is the key to obtain high selectivity of arylamines. However, the underlying reaction mechanism of route regulation is uncertain owing to a lack of direct in situ spectral evidence of the dynamic transformation of intermediate species during the reaction process. In this work, by using in situ surface-enhanced Raman spectroscopy (SERS), we have employed 13 nm Au100−xCux nanoparticles (NPs) deposited on a SERS-active 120 nm Au core to detect and track the dynamic transformation of intermediate species of hydrogenation of para-nitrothiophenol (p-NTP) into para-aminthiophenol (p-ATP). Direct spectroscopic evidence demonstrates that Au100 NPs exhibited a coupling route with the in situ detection of the Raman signal assigned to coupling product p,p′-dimercaptoazobenzene (p,p′-DMAB). However, Au67Cu33 NPs displayed a direct route without the detection of p,p′-DMAB. The combination of X-ray photoelectron spectroscopy (XPS) and density functional theory (DFT) calculations reveals that Cu doping can favor the formation of active Cu–H species owing to the electron transfer from Au to Cu, which can promote the formation of phenylhydroxylamine (PhNHOH*) and favor the occurrence of the direct route on Au67Cu33 NPs. Our study provides direct spectral evidence demonstrating the critical role of Cu in route regulation for the nitroaromatic hydrogenation reaction at a molecular level and clarifies the route regulation mechanism. The results have significant implications for revealing multimetallic alloy nanocatalyst mediated reaction mechanisms and help to guide the rational design of multimetallic alloy catalysts for catalytic hydrogenation reactions.

(HAuCl 4 ) was rapidly added to the boiling solution. The color of the mixture solution changed from pale yellow to bluish-gray and then to burgundy. Then the solution temperature was slowly cooled to 90 °C under continuous stirring. At the moment, 1 mL 25 mM HAuCl 4 solution was added to the above system quickly and kept for 30 min. After 30 min, 1 mL 25 mM HAuCl 4 solution was added to the above system quickly and kept for another 30 min. Then, 55 mL of the above solution was named as g x (x=05) and was used as a seed solution to obtain Au NPs of a larger size. 53 mL water and 2 mL 60 mM sodium citrate aqueous were added to the 55 mL seed solution. Then, 1 mL HAuCl 4 solution (25 mM) was injected into the above solution quickly, 30 min later, 1 mL HAuCl 4 solution (25 mM) was injected into the above solution quickly and kept for another 30 min. The above steps were repeated five times until the diameter of the final product (g x ) was about 120 nm.

Synthesis of Au@SiO 2 substrate.
The Au@SiO 2 was prepared based on the literature report. 2 Typically, 0.4 mL 1 mM (3-aminopropyl) trimethoxysilane (APTMS) was added to the as-prepared 120 nm Au NPs at room temperature under vigorous stirring for 3 min. Then, 3.2 mL 0.54% NaSiO 3 aqueous solution was introduced to the above solution under intense stirring at room temperature for 3 min. The mixture solution was heated to 98 °C for 20 min to accelerate the coating of the ~3 nm SiO 2 shell on the surface of 120 nm Au NPs.
After that, the mixture was cooled rapidly in an ice bath to prevent the further growth of SiO 2 shell. Eventually, the Au@SiO 2 substrate was obtained by centrifugation at 4000 rpm for 5 min. Finally, the obtained Au@SiO 2 were redispersed in 1 mL ethanol for further experiments.

Synthesis of 13 nm Au 100 NPs.
The 13 nm Au 100 NPs were prepared by the sodium citrate reduction method. 3 Firstly, 100 mL 1 mM HAuCl 4 aqueous solution was heated to boiling under constantly stirring. After that, 10 mL 38. nm Au 90 Cu 10 is similar with that of 13 nm Au 67 Cu 33 . The difference is that the input of acetyl copper is 28 mg.

Computational method.
We have employed the Vienna Ab Initio Package (VASP) 8,9 to perform all spinpolarization DFT calculations within the generalized gradient approximation (GGA) using the Perdew-Burke-Ernzerhof (PBE) formulation. 10 In addition, the adsorption model was optimized by Newton's method of second order convergence (IBRION=1 and POTIM=0.2). We have chosen the projected augmented wave (PAW) potentials 11,12 to describe the ionic cores and take valence electrons into account using a planewave basis set with a kinetic energy cutoff of 520 eV. Partial occupancies of the Kohn−Sham orbits were allowed to use the Gaussian smearing method and a width of 0.05 eV. The electronic energy was considered self-consistent when the energy change was smaller than 10 −5 eV. Geometry optimization was considered convergent when the energy change was smaller than 0.
where ΔE is the difference between the total energy, ΔEZPE and ΔS are the differences in the zero-point energy and the change of entropy, T is the temperature (T = 300 K in this work).

Calculated procedure for SNR of SERS signal for different catalysts.
The SNR of SERS signal for different catalysts were calculated according to the literature report. 13 Signal-to-Noise-Ratio (SNR) is defined by:

SNR=10lg
Where P signal and P noise are the power of signal and the power of noise.
In order to calculate the SNRs of SERS signal for different catalysts, five different Raman spectra were obtained on the same catalyst without the addition of NaBH 4 .
Following is the detailed procedure: i. The signal power (P signal ) was calculated according to the equation: P signal = . The value of I signal of the five different Raman spectra is obtained by ii. The noise power (P noise ) was calculated according to the equation: P noise = . In order to obtain I noise of the five different Raman spectra, the average iii. The SNR of each catalyst is calculated by SNR=10lg( ) The result is given in Table S1.      Figure S6. In situ SERS spectra of p-NTP on Au@SiO 2 @Au 100 without NaBH 4 . Figure S7. The SERS intensities of (a) background and (b) the initial Raman spectra of p-NTP molecule on the three types Au@SiO 2 @Au x Cu x substrates.
The initial surface coverage of p-NTP was proportional to the initial Raman intensity of the peak at 1570 cm −1 for p-NTP molecules on the three different catalysts.
The result shows that the initial Raman intensity of the peak at 1570 cm −1 for p-NTP molecules is similar, indicating that the initial surface coverage of p-NTP is similar.